Industrial Production: Electric and Gas Utilities

In this article, we use a dataset from Fred Economic Data. In particular, we use Industrial Production (electric and gas utilities) from 1939-01-01 until Now (the last available dataset). This dataset can be obtained here.

Dataset Notes

The industrial production (IP) index measures the real output of all relevant establishments located in the United States, regardless of their ownership, but not those located in U.S. territories.

First, we can use seasonal decompose function (read more about this function here to analyze the Data.

Furthermore, here we use the auto_arima function that Automatically discovers the optimal order for an ARIMA model for the sake of predictions. Ideally, we are looking for a model that fits the data and has the lost AIC score, AIC stands for the Akaike information criterion. the AIC score gives provides a way to measure the goodness-of-fit of your model.

Train and Set Tests

We can consider the data from the current year as the Test set and the rest of the data as the train set. That is

Now, let's train the model

Predications


References

  1. Fred, Industrial Production: Electric and gas utilities, https://fred.stlouisfed.org/series/IPG2211A2N
  2. Statsmodels User Guide
  3. ARIMA estimators for Pytho
  4. Akaike information criterion Wikipedia page
  5. Autoregressive integrated moving average Wikipedia page